module: unzip (#642)

* Specify more guidelines on input channels

* Linting

* Updates based on code review

* Update README.md

* Fix broken sentence

* Add unzip module

* Remove missing TODOs update mtea

* Apply changes after code-review from @grst

* Account for user trying to supply two input archives

* Remove debugging test

* Update modules/unzip/main.nf

Co-authored-by: Jose Espinosa-Carrasco <kadomu@gmail.com>

* Correct output path

Co-authored-by: Jose Espinosa-Carrasco <kadomu@gmail.com>

modules/unzip/functions.nf

@@ -0,0 +1,68 @@
+//
+//  Utility functions used in nf-core DSL2 module files
+//
+
+//
+// Extract name of software tool from process name using $task.process
+//
+def getSoftwareName(task_process) {
+    return task_process.tokenize(':')[-1].tokenize('_')[0].toLowerCase()
+}
+
+//
+// Function to initialise default values and to generate a Groovy Map of available options for nf-core modules
+//
+def initOptions(Map args) {
+    def Map options = [:]
+    options.args            = args.args ?: ''
+    options.args2           = args.args2 ?: ''
+    options.args3           = args.args3 ?: ''
+    options.publish_by_meta = args.publish_by_meta ?: []
+    options.publish_dir     = args.publish_dir ?: ''
+    options.publish_files   = args.publish_files
+    options.suffix          = args.suffix ?: ''
+    return options
+}
+
+//
+// Tidy up and join elements of a list to return a path string
+//
+def getPathFromList(path_list) {
+    def paths = path_list.findAll { item -> !item?.trim().isEmpty() }      // Remove empty entries
+    paths     = paths.collect { it.trim().replaceAll("^[/]+|[/]+\$", "") } // Trim whitespace and trailing slashes
+    return paths.join('/')
+}
+
+//
+// Function to save/publish module results
+//
+def saveFiles(Map args) {
+    if (!args.filename.endsWith('.version.txt')) {
+        def ioptions  = initOptions(args.options)
+        def path_list = [ ioptions.publish_dir ?: args.publish_dir ]
+        if (ioptions.publish_by_meta) {
+            def key_list = ioptions.publish_by_meta instanceof List ? ioptions.publish_by_meta : args.publish_by_meta
+            for (key in key_list) {
+                if (args.meta && key instanceof String) {
+                    def path = key
+                    if (args.meta.containsKey(key)) {
+                        path = args.meta[key] instanceof Boolean ? "${key}_${args.meta[key]}".toString() : args.meta[key]
+                    }
+                    path = path instanceof String ? path : ''
+                    path_list.add(path)
+                }
+            }
+        }
+        if (ioptions.publish_files instanceof Map) {
+            for (ext in ioptions.publish_files) {
+                if (args.filename.endsWith(ext.key)) {
+                    def ext_list = path_list.collect()
+                    ext_list.add(ext.value)
+                    return "${getPathFromList(ext_list)}/$args.filename"
+                }
+            }
+        } else if (ioptions.publish_files == null) {
+            return "${getPathFromList(path_list)}/$args.filename"
+        }
+    }
+}

modules/unzip/main.nf

@@ -0,0 +1,43 @@
+// Import generic module functions
+include { initOptions; saveFiles; getSoftwareName } from './functions'
+
+params.options = [:]
+options        = initOptions(params.options)
+
+process UNZIP {
+    tag "$archive"
+    label 'process_low'
+    publishDir "${params.outdir}",
+        mode: params.publish_dir_mode,
+        saveAs: { filename -> saveFiles(filename:filename, options:params.options, publish_dir:getSoftwareName(task.process), meta:[:], publish_by_meta:[]) }
+
+
+    conda (params.enable_conda ? "bioconda::p7zip=15.09" : null)
+    if (workflow.containerEngine == 'singularity' && !params.singularity_pull_docker_container) {
+        container "https://depot.galaxyproject.org/singularity/p7zip:15.09--h2d50403_4"
+    } else {
+        container "quay.io/biocontainers/p7zip:15.09--h2d50403_4"
+    }
+
+    input:
+    path archive
+
+    output:
+    path "${archive.baseName}/" , emit: unzipped_archive
+    path "*.version.txt"     , emit: version
+
+    script:
+    def software = getSoftwareName(task.process)
+
+    if ( archive instanceof List && archive.name.size > 1 ) { exit 1, "[UNZIP] error: 7za only accepts a single archive as input. Please check module input." }
+
+    """
+    7za \\
+        e \\
+        -o"${archive.baseName}"/ \\
+        $options.args \\
+        $archive
+
+    echo \$(7za --help) | grep Version | sed 's/.*p7zip Version//; s/(.*//' 1> ${software}.version.txt
+    """
+}

modules/unzip/meta.yml

@@ -0,0 +1,31 @@
+name: unzip
+description: Unzip ZIP archive files
+keywords:
+  - unzip
+  - decompression
+tools:
+  - unzip:
+      description: p7zip is a quick port of 7z.exe and 7za.exe (command line version of 7zip, see www.7-zip.org) for Unix.
+      homepage: https://sourceforge.net/projects/p7zip/
+      documentation: https://sourceforge.net/projects/p7zip/
+      tool_dev_url: https://sourceforge.net/projects/p7zip"
+      licence: "GNU LPGL"
+
+input:
+  - archive:
+      type: file
+      description: ZIP file
+      pattern: "*.zip"
+
+output:
+  - version:
+      type: file
+      description: File or directory of decompressed archive
+      pattern: "*.{version.txt}"
+  - unzipped_archive:
+      type: directory
+      description: Directory contents of the unzipped archive
+      pattern: '${archive.baseName}/'
+
+authors:
+  - "@jfy133"

tests/config/pytest_modules.yml

@@ -835,6 +835,10 @@ untar:
   - modules/untar/**
   - tests/modules/untar/**
 
+unzip:
+  - modules/unzip/**
+  - tests/modules/unzip/**
+
 variantbam:
   - modules/variantbam/**
   - tests/modules/variantbam/**

tests/config/test_data.config

@@ -7,6 +7,7 @@ params {
             'genome' {
                 genome_fasta                                   = "${test_data_dir}/genomics/sarscov2/genome/genome.fasta"
                 genome_fasta_fai                               = "${test_data_dir}/genomics/sarscov2/genome/genome.fasta.fai"
+                genome_fasta_zip                               = "${test_data_dir}/genomics/sarscov2/genome/genome.fasta.zip"
                 genome_dict                                    = "${test_data_dir}/genomics/sarscov2/genome/genome.dict"
                 genome_gff3                                    = "${test_data_dir}/genomics/sarscov2/genome/genome.gff3"
                 genome_gff3_gz                                 = "${test_data_dir}/genomics/sarscov2/genome/genome.gff3.gz"

tests/modules/unzip/main.nf

@@ -0,0 +1,12 @@
+#!/usr/bin/env nextflow
+
+nextflow.enable.dsl = 2
+
+include { UNZIP } from '../../../modules/unzip/main.nf' addParams( options: [:] )
+
+workflow test_unzip {
+
+    archive = file(params.test_data['sarscov2']['genome']['genome_fasta_zip'], checkIfExists: true)
+
+    UNZIP ( archive )
+}

tests/modules/unzip/test.yml

@@ -0,0 +1,7 @@
+- name: unzip
+  command: nextflow run ./tests/modules/unzip -entry test_unzip -c tests/config/nextflow.config
+  tags:
+    - unzip
+  files:
+    - path: output/unzip/genome.fasta/genome.fasta
+      md5sum: 6e9fe4042a72f2345f644f239272b7e6